[01:12] dammit, the inspector isn't a real view and doesn't have YUI events available to it. It can't this.fire or inspector.on() [02:22] I hate this test :T AssertionError: expected 300 to equal 298.552734375 [02:22] hah, JS match...you're drunk [02:22] math that is [03:22] evening [03:22] party [03:25] hows it going? [03:25] bed time, branch is pushed for the night === rogpeppe2 is now known as rogpeppe [13:08] benji: pushed another update to that branch per hatch's comments. [13:08] rick_h_: cool, I'll take a look === gary_poster|away is now known as gary_poster [13:23] rick_h_: do you have any tips on diffing your branch and mine? git diff auto-open-close..qa-auto-open-rick (with two or three dots) shows lots of things I wouldn't expect [13:52] benji: you should be able to diff hashes in line. Since I started with your branch [13:52] benji: git checkout qa-auto-open-rick [13:52] git log [13:52] find the merge point [13:53] and then diff $hash..HEAD? [13:53] rick_h_: I'll give that a try. I was hoping for something more direct. [13:56] benji: https://github.com/mitechie/juju-gui/compare/benji-york:auto-open-close...mitechie:auto-open-close [13:56] benji: hmm, crap we have different develop bases which is why things are there you don't expect [13:57] benji: so the key would be to get the common base point by updating develop and getting that into your branch [14:40] rick_h_ so is PR 91 still a WIP? I was looking at it last night and assumed so... [14:40] hatch: yea, it's got two failing tests still [14:40] hatch: but it was something to see if I could get the origin idea working based on benji's branch [14:41] original* [14:41] cool lemme know when she's ready [14:41] hatch: will do, thanks for the pre-review. It made me clean up a little bit of what I should have [14:45] rick_h_: I'm looking at the two test failures [14:46] benji: it should just be another round of bootstraping all the stuff required to create an Inspector instance. None of the current tests run createServiceInspector or whatever that func is [14:46] benji: so that has to be called before testing the events [14:46] rick_h_: ok [14:46] benji: so that you have a this.inspector.viewletManager [14:46] * rick_h_ thinks that's what it was thinking back to when he gave up last night === gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [14:55] ubuntu mouse stopping work [14:55] that is, ubuntu mouse handling broke again and I had to log out/in [14:56] hey frankban, I spoke with jamespage about Ubuntu packaging and he requested that we put quickstart on PyPI so he has official releases to work with [14:56] is that something you could whip up quickly? [14:56] gary_poster: sounds good. I think so, what Pypi account am I supposed to use? [14:57] frankban: I suggest yours, and then add everyone else on GUI with a PyPI acct as admin [14:57] frankban: you can use yours but I think you can grant access to other pypi users [14:57] gary_poster, rick_h_ cool [14:57] frankban: mitechie is my username here [15:00] frankban: i see that quickstart starts an ssh-agent whether one is running or not. couldn't we check to see if one is alive and happy before clobbering the existing one? [15:01] frankban on PyPI I am "gary" and benji is "benji". I don't see bac or Makyo on a quick glance. [15:01] actually maybe you don't need to add me :-P [15:01] bac: Makyo knows the details for SSH keys handling [15:01] creating a new ssh-agent every run means i have to re-enter my passphase. [15:02] gary_poster: I'll add you! :-/ [15:02] :-) k [15:04] frankban: add bac please [15:26] the best part about OSX is coming across bugs, googling them, then seeing that people have had the same issue for 4 years [15:27] gary_poster: could you please take a quick look at https://codereview.appspot.com/51790045 ? (quickstart update for PyPI) [15:27] on it [15:27] thanks [15:28] frankban: LGTM'd :-) [15:31] gary_poster: thank you [15:32] np! [15:43] jujugui https://pypi.python.org/pypi/juju-quickstart [15:43] bac: no "bac" user found on PyPI [15:44] hatch, Makyo: what's your PyPI user name? [15:44] awesome thanks frankban [15:44] wtpi is pypi [15:44] :P [15:45] hatch: please go to https://pypi.python.org/pypi, register, give me your username and then enjoy the cheese [15:46] lol ok [15:47] sweet I got 'hatch' as a username [15:47] that never happens [15:50] jujugui call in10 === mhall119_ is now known as mhall119 [15:58] jujgui call in 2 [15:58] jujugui [16:01] Makyo: pingy [16:02] Sorry, SSO. [16:11] frankban: try bradcrittenden [16:11] for pypi [16:11] bac: it worked [16:11] \o/ [16:20] frankban I'll ping you after my call :) [16:20] hatch: ok [16:21] gary_poster: can you invite me or link me to the meeting please? [16:24] gary_poster do you know PostreSQL? Linkedin wants to know...lol [16:24] hatch: but of course! ;-) [16:26] annnnd signing out to log back into linked in in another year [16:27] hazmat: working on the feedback branch for the deployer, I should be able to propose it tomorrow morning [16:32] frankban, cool, thanks [16:32] hatch, pypi == npm [16:32] registry [16:32] mp [16:33] not sure which website is worse pypi or npm though [16:33] not to be confused with pypy [16:33] or with pie pi [16:34] the later of which is yummy and mind expanding :-) http://upload.wikimedia.org/wikipedia/commons/d/d4/Pi_pie2.jpg [16:38] http://www.youtube.com/watch?v=jG7vhMMXagQ [17:18] frankban still around? [17:18] I have no idea when your EOD is :) [17:19] rick_h_ are you available to guide me through this testing business? [17:19] python testing that is [17:19] hatch: on a call but can in a bit [17:19] ok cool np, ping when ready [17:19] might trade some help [17:20] right now my help is selling at a 2:1 help unit ratio [17:20] market conditions... [17:20] :P [17:22] hatch: I am here, EOD in 40 mins [17:22] oh ok cool, do you have time to guide me through the testing now? [17:22] hatch: sure [17:23] https://plus.google.com/hangouts/_/7ecpjd60csb16idpl8krnvb4es?authuser=1&hl=en [17:25] ^ frankban [17:25] hatch: joining [17:43] frankban: i get the following while doing QA on my branch. https://pastebin.canonical.com/103823/ [17:44] bac: one minute, on call [17:44] frankban: quickstart terminates with juju-quickstart: error: bad API server response: state watcher was stopped [17:44] frankban: ok. just curious if it looked familiar [17:48] bac: yes, that's supposed to be fixed in 0.17.1 [18:04] yeah... [18:12] jujugui: per frankban's message, anyone know if juju 0.17.1 is packaged anywhere or do i need to build by hand? [18:12] bac it's released already [18:12] yup [18:12] i don't see it on trusty [18:12] hmm the email said it was available on trusty [18:13] bac: it's released today [18:13] at least I thought it did [18:13] juju-core 1.17.1 is available for trusty and backported to earlier [18:13] series in the following PPA: [18:13] https://launchpad.net/~juju/+archive/devel [18:13] jujugui pita review, read the notes. Had to skip the notifier tests atm. It needs rework and adding as a new card. https://github.com/juju/juju-gui/pull/89 [18:13] i'll take one [18:14] hatch: thanks [18:14] how many do you need rick_h_ [18:14] gary_poster: 2 please. thought it's mostly mechanical so fine with just one [18:14] gary_poster: the main thing is the heads up on the notifier tests. [18:14] rick_h_: ok, I'll leave it alone unless requested then :-) [18:15] thanks hatch. that ppa got disabled on upgrade to trusty and i failed to re-enable it [18:16] ahh [18:25] rick_h_ 44 files changed? lol [18:25] hatch: well, makeContainer is popular :) [18:25] jujugui: i need to be afk for the rest of the day. ttyl. [18:25] bac: c-ya [18:26] yay for the end of conference calls [18:27] rick_h_ you're gona hate me... [18:28] hatch: then stop [18:28] lol [18:28] hatch: I'm already very angry after chasing down the stupid notifier test bugs [18:28] :OP [18:28] :P [18:29] rick_h_ what's with the onboarding.js changes [18:29] hatch: it didn't clean up after itself [18:29] now it does [18:29] lots of tests are like this unfortunately. Creating objects that never die [18:29] well when it's container gets destroyed it would remove it's children [18:29] which is all you're doing there....no? [18:30] no, it's got events bound and the tests never call destory on the object [18:30] * rick_h_ pulls it back up [18:31] ohh nm empty removes the events too [18:31] hatch: oh sorry, so this is that the object doesn't have a destructor itself. see the test_onboarding file that goes with it [18:31] matching changes [18:31] * rick_h_ kicks self for missing this during review/helping getting onboarding landed. [18:34] hatch: so what am I going to hate you for? [18:35] I am trying to figure out a way to avoid having to pass 'this' into makeContainer every time [18:35] hatch: good luck, it's stupid mocha scopes [18:35] hatch: everything has to be in the hook context to be visible unless I stick it on window. which I almost did...but that feels evil and wrong [18:36] that's irritating [18:36] and there's no way to globally rebind the makeContainer function from the juju test utils scope since each test has it's own Hook context [18:36] yeah I just noticed that [18:36] how fail is that :/ [18:36] you'd have to loop throgh all the Suites, and down in there go through each Hook context of each, and they can nest and ... [18:37] I told you, I'm very cranky with Mocha, etc atm. This was a much bigger pita than it should be, but oh well. [18:37] hatch: but trust me, I tried to make it not suck, if you find a better way I'll thank you for the insight and move alnog [18:37] lol [18:37] well their scoping is totally wako [18:37] yes...yes it is [18:37] it doesn't make any sense [18:38] sure it does, it's how the damn things works. It laods all the files, sticks all the tests in their own Hook contexts, and basically loops them [18:38] in a nested pita tree of SAuites [18:38] Suites [18:38] lol well ok it makes sense, it's just not very friendly [18:38] but it makes it darn hard to monkey patch things [18:38] or even trace/debug things [18:39] anyway, going to walk away and nuke some lunch. let me know if you have any ideas on improving. I'm definitely for it, but in the end I think that the api of needing to pass this isn't bad. I even catch/report an error that is hopefully clear [18:50] hatch: the notifier stuff is noted in the pull request [18:51] yeah I replied again, guessing it didn't send another email for that :) [18:51] card added to fix it, but will be second card [18:51] ah, well email checks every so often so might not have it yet [18:54] rick_h_ what is the model of your laptop? [18:54] hatch: x240 [18:54] rerr 230 [18:58] cool [18:58] guy is just asking what people use for linux laptop hardware [19:34] Makyo: heads up, test branch landed so watch out for conflicts [19:34] >:/ [19:34] benji: ^ [19:34] thanks [19:35] russian roulette, touched 44 files and was one of them yours?! bwuhahaha [19:36] rick_h_: I think I have the test setup right for those two failing tests, but... the event the tests fire doesn't exist anywhere else in the codebase [19:36] rick_h_: oh! it's a typo [19:36] benji: yay [19:37] "inpsector" [19:37] doh [19:37] I kept mixing up the case last night [19:37] Takeover vs TakeOver [19:37] tab completion [19:37] and such [19:45] nice, jcastro giving quickstart some <3 http://discourse.ubuntu.com/t/mangling-config-files-by-hand-pshawwww/1438 [19:45] purdy screenshots [19:45] I _love_ the "has errors" bit [19:45] I totally didn't notice that the first time around [19:47] jcastro: yea, hopefully we can get it into main/universe https://launchpad.net/bugs/1273865 and make for an awesome getting started experience [19:47] I'd imagine it's gotta be cool for your on site classes [19:47] oh hey, does the server team know about that bug? [19:47] jcastro: yea, we pinged james page on it and we're working on it [19:48] they're handling juju but I think we should make clear that juju without quickstart will be kind of lame [19:48] ok, nice! [19:48] rick_h_, I showed it off at a class first actually [20:13] stupid power button [20:28] jujugui why is minor < 17 false when minor is 16 in python? [20:28] does maths work in reverse in python? lol [20:29] hatch: code? maybe the types are off [20:29] is minor a string? [20:29] hatch [20:29] >>> "16" < 17 [20:29] False [20:29] >>> 16 < 17 [20:29] True [20:29] https://gist.github.com/hatched/8696341 [20:30] hatch: right, a version might be 16a [20:30] so it's a string ootb [20:30] strings to ints [20:30] one of imporvements in 3.3: [20:30] >>> "16" < 17 [20:30] Traceback (most recent call last): [20:30] File "", line 1, in [20:30] TypeError: unorderable types: str() < int() [20:30] yay [20:30] >>> "1" >= 1 [20:30] True [20:30] I thought js was messed up lol [20:31] >>> ord("1") [20:31] 49 [20:31] 49 > 1 :) [20:32] welcome to ascii or unicode as the python version you're using might be [20:32] 2.7.3 [20:32] but yes, this is one of the warts that py 3 was made to fix [20:32] whoa, chrome update looks so much nicer yay! [20:32] ok so int(minor) should fix ? [20:32] sorry I mean [20:33] is using int() the 'proper' way [20:34] yes [20:34] cool [20:34] and people rag on JS for having oddities haha [20:34] error catching on conversion is also nice [20:35] hatch: yea, toss a test for a version with a character in it and make sure it's caught and handled ok. [20:35] it should have thrown by this point so that won't be necessary [20:35] hatch: heh, the issue is JS magically casts it for you using its own rules and that leads to wat videos [20:35] will do [20:35] hey THIS would have been a wat video too! [20:35] lol [20:36] well how resilient will I need to make this method? Is Juju switching form semver any time soon? haha [20:44] gary_poster, retcon on the feature flag :( I forgot about the remove relation dialog modifications I made. [20:44] Would be >15min [20:45] We just have a pre-indicator now. [20:48] Makyo: ack, good 'nuff [20:51] * gary_poster has to go and work offline. I will be on email. === gary_poster is now known as gary_poster|away [21:00] I think the python linter is more picky than the js one [21:04] jujugui does anyone know if quickstart uses lbox to propose? [21:04] hatch: yes. you can tell by the .lbox files [21:04] hatch: I believe it does. If it has a .lbox file it almost certainly does [21:06] thanks, guess this vm is missing lbox and the gophers ppa heh [21:07] may apt-get have mercy on your soul [21:07] Shoot. [21:07] This also affects bundle visualizations. [21:07] umm this vm doesn't have a GUI [21:07] lol [21:08] how do I auth with launchpad [21:08] Uh, crap, I forgot. [21:08] Makyo didn't you run into this issue some time? [21:08] install links2 and update-alternatives? [21:08] the lbox w/o a gui? [21:10] rick_h_ that really the only way? [21:11] hatch: doesn't it spit out a LP url for lplib auth? if so, you could visit it in another browser. no? [21:11] bac it doesn't [21:12] hmm [21:12] at least here it's not [21:12] hatch: you can always just do a LP MP [21:12] old skool [21:12] that will skip reitveld though no? [21:12] yes [21:12] yeah I think i'd much prefer someone to comment on my first mostly-solo python branch hah [21:13] hatch: still can do reviews on a launchpad merge proposal [21:13] in fact, we insist [21:13] ohh I didn't know that [21:13] in that case... [21:14] push your branch, go to it in the browser of your choice, select something that smells like "propose for merging" [21:14] yup typing it out now [21:15] jujugui lf quickstart review and qa https://code.launchpad.net/~hatch/juju-quickstart/no-more-sudo/+merge/203846 [21:25] http://shop.oreilly.com/product/0636920028154.do wow 1600 pages for learn python book hah [21:46] Oh hush. [21:47] You're nearing 'evangelism', but you work for a company that lists python as a core technology :P [21:48] (Though juju in Node would certainly be interesting :D ) [21:49] haha [21:49] everyone knows node is faster than python anyways [21:49] * hatch ducks and runs [21:50] seriously though, 1600 pages is a lot for any book haha [22:06] Morning [22:06] hey huwshimi [22:09] Makyo definitely like the emoticon in your PR [22:09] are those fish? [22:10] Ahahaha! [22:10] They're pennants. [22:10] Our feature flags are now actual flags. [22:10] Boo, build failed. Investigating. [22:11] lol [22:11] DAMNIT. I ran make prep >:/ [22:12] one commit ago. [22:12] hmm I am running juju 1.17.1 and it's requiring sudo for local [22:17] hatch: This is the correct way to check if the viewport has changed right? Y.on('windowresize', function(e) {... [22:17] just resize [22:17] Y.on('resize', function(e) { ... [22:19] huwshimi although there is a windowresize which normalizes the resize event across browsers [22:19] event-resize I think [22:19] hatch: Hmm.... neither seem to be firing [22:19] try Y.config.doc.on('resize' .. [22:19] or Y.config.win [22:21] sec looking [22:22] huwshimi you can probably just get away with Y.one('window').on('resize', ... [22:22] hatch: config.doc and config.win don't exist [22:22] Y.cfg? maybe [22:22] crap now I gota go look [22:23] yeah it's Y.config.doc [22:23] and Y.config.wion [22:23] win* [22:23] Y.one('window').on('resize' works [22:23] so there really is nothing in Y.config.doc? [22:23] in jujucharms.com there is [22:24] but it's fine the Y.one('window') is ok, Y.one(Y.config.win) would be better though :) [22:24] if it actually exists [22:25] Y.config.win/doc are supposed to be referenced to the 'actual' cached doc and windowreference [22:25] Ah, I mis-read the error: Uncaught TypeError: Object # has no method 'on' [22:25] ohh ok [22:25] so they changed that [22:25] yeah Y.one('Y.config.win).on('resize', ... [22:26] yeah [22:26] it used to be a Y.Node, they must have changed that for perf [22:26] Ah ok [22:27] hatch: Strangely though that seems to fire twice on every resize. [22:28] put a debugger before you attach your event, it's likely that you are attaching it twice [22:28] resize also fires periodically as the user is resizing so you 'may' also be seeing that [22:30] hatch: It looks like the attachment only happens once, but the event fires twice [22:30] hmm [22:31] and the event objects passed into the callbacks are the same? [22:32] hatch: I wonder if we only want to show this message once in a session? [22:33] hatch: zooming in the browser fires a viewport resize event so it'll pop up a lot of messages while you're zooming [22:35] ok load in the event-resize module (add it to the requires) and then listen on windowresize [22:35] that debounces the resize events [22:36] hatch: What does that do? [22:36] sorry debouncing means that every time something is fired it resets a time [22:36] r [22:36] once that timer runs out it fires the event [22:37] hatch: Do you happen to know what that time is? [22:37] Actually we might not want a timer here [22:38] 100ms http://yuilibrary.com/yui/docs/api/files/event_js_resize-window.js.html#l27 [22:38] For example if you're zooming back in after a few minutes you probably don't want to see the message again [22:38] ohh right, I don't actually know what you're working on :) [22:39] :)